Multiple resolution video encoding and decoding is described, inter alia, in Holcomb et al., “Multi-resolution Video Coding and Decoding,” U.S. patent application Ser. No. 10/644,258, filed Aug. 19, 2003 (the disclosure of which is hereby incorporated herein by reference) [hereafter the '258 patent application]. In multi-resolution video encoding, the video encoder may adaptively change or lower the coded size of video frames within a video sequence relative to the displayed video size so as to reduce blocking artifacts at low bitrates (among other reasons). For video frames coded at lower resolution, the frame is down-sampled at the encoder to permit coding at a lower resolution, then up-sampled at the decoder to maintain a consistent display resolution. One implementation of the multiple resolution video encoding and decoding techniques described in this patent application is incorporated in simple and main coding profiles of the Windows Media Video version 9 (WMV-9) video codec distributed and licensed by Microsoft Corporation of Redmond, Wash., USA.
In the WMV-9 video codec's simple and main coding profiles, the video encoder communicates the coding resolution to the decoder using elements of the codec's bit stream syntax. In particular, the bit stream syntax includes a multi-resolution flag syntax element (labeled “MULTIRES”) at the sequence header that indicates whether or not there could be resolution changes within the bit stream, and also includes a two-bit picture resolution syntax element (labeled “RESPIC”) at the picture header to signal a reduction in resolution by a factor of two (i.e., reduction to one-half the display resolution) in the horizontal or vertical dimensions of the picture. The sequence header was carried at the level of the file container (e.g., the “ASF” format file container), externally from the encoded video bit stream contained therein.